The Overall Objective of Project 1 is to investigate both interactive and autonomous (computer optimization) techniques for designing conformational radiation therapy treatment plans based on the new computer-controlled radiotherapy (CCRT) treatment technology. To improve interactive design, we will investigate the use of higher level beam objects like arcs, cones, and curved paths that can be delivered dynamically (or approximated with a series of "segmental" beams). We will also investigate the use of graphical volumetric analysis techniques to aid in the placement of beams. To further improve planned dose distributions, we propose to integrate these techniques with automated optimization based on the simulated annealing algorithm combined with the use of realistic dose response scoring functions. We will investigate techniques for improving the rate of convergence of using adaptive or "self-regulating" cooling schedules and by incorporating techniques from deterministic optimization algorithms. To create and demonstrate that optimal plans can be clinically realized, this project will rely on the other projects to provide data on calculational and delivery accuracy and constraints (Project 2), on geometrical accuracies (Project 3) and on dose response data (Project 4). In turn, these other projects will benefit from the identification of the most critical factors affecting the optimization of conformal treatments. An overall assessment of the potential improvements possible as well as the relative value of the various plan optimization techniques will be analyzed through a series of comparative planning studies (using clinical cases which are being treated on the conformal therapy protocols in Project 4). Comparisons will be made between the results of conventional planning and the results achievable with the new conformal planning techniques developed in this project.